GSN: A Graph-Structured Network for Multi-Party Dialogues
This addresses the challenge of dialogue response generation for multi-party conversations, which is incremental as it extends existing sequential models to graph structures.
The paper tackles the problem of modeling multi-party dialogues where utterances can occur in parallel, by generalizing sequence-based models to a Graph-Structured Network (GSN), and shows that GSN significantly outperforms existing sequence-based models.
Existing neural models for dialogue response generation assume that utterances are sequentially organized. However, many real-world dialogues involve multiple interlocutors (i.e., multi-party dialogues), where the assumption does not hold as utterances from different interlocutors can occur "in parallel." This paper generalizes existing sequence-based models to a Graph-Structured neural Network (GSN) for dialogue modeling. The core of GSN is a graph-based encoder that can model the information flow along the graph-structured dialogues (two-party sequential dialogues are a special case). Experimental results show that GSN significantly outperforms existing sequence-based models.